The inventive subject matter relates to electric power systems and methods of operating the same and, more particularly, to uninterruptible power supply systems and methods of operating the same.
Conventional electrical power systems used in commercial applications, such as data centers, typically include various types of devices, such as switchgear units, transformers, power distribution units (PDUs) and uninterruptible power supplies (UPSs). These are typically single purpose units and are selected and interconnected to suit a particular application. For example, a conventional UPS used in data center power system may include a cabinet or a modular arrangement of cabinets that has a relatively fixed topology, e.g., a particular interconnection of rectifier, inverter and battery converter circuits that is designed to provide a power output from a primary source, such as a utility source, and a secondary source, such as a battery. The primary and secondary power sources may be connected to the UPS and other power network components using switchgear and other network components that typically are selected for the particular application. Such units may be difficult to integrate with one another and with other devices. Some UPS systems may have modular construction in order to facilitate capacity scaling and replacement in case of failure, but the modules used in such systems are typically configured for use in a fixed arrangement.
Large data centers have proliferated with the advent of web services and cloud computing. Some newer large data centers occupy millions of square feet and house hundreds of thousands of servers. These centers may have varying power requirements. For example, a data center may host both fault-tolerant applications, such as social media and video services, and fault-intolerant applications, such as financial applications. Energy consumption is a major concern for such facilities, as some facilities are approaching the 100 MW level, where even a few percentage points of lost efficiency can translate into significant expense. It may be desirable to power servers hosting fault-intolerant applications using highly reliable systems, such as on-line UPSs. However, running fault-tolerant applications on servers with a highly-reliable UPS may be relatively inefficient. Various solutions for providing power in data centers with relatively high efficiency and redundancy are proposed, for example, in U.S. Pat. No. 7,886,173 to Krieger et al., U.S. Pat. No. 7,560,831 to Whiffed et al. and U.S. Pat. No. 8,344,546 to Sarti. These solutions, however, may be relatively inflexible and may not provide a sufficient breadth of capabilities.